Interview with Bryan Valentini: A Graduate of R-Beginner & R-Intermediate Programs
Azeem: Why don't we start by you telling us a bit about your background.
Bryan: My background is in software. I graduated from Carnegie Mellon where I studied computer science and HCI (Human Computer Interaction). HCI is basically studying how to best design interfaces and computer systems in general that match the way humans think. I got my start there, and I gravitated towards visualization. That's where I went after college. I worked in a small start up in Pittsburgh that was building a very interesting mapping and visualization product. I got my chops there. I recently transitioned to the start up world here in NYC where I learned about digital advertising and getting involved in the "internet of things" and building prototypes and hardware
Azeem: What made you decide to take classes at NYC Data Science Academy?
Bryan: I really wanted to start solving the problem of having visualizations build themselves, rather than having to manually specify the data and mapping to the symbols that show on a chart. I wanted a system to be able to figure it out itself, not only in terms of visualizing data, but finding trends, looking for anomalies, finding the outliers and trying to bring that to the human's attention. That involved aspects of machine learning and statistics, and the more I got into it, the more interesting I found it. Like any data geek - when you start building data systems you go from wanting to build the low level machinery to the higher level machinery; you start saying, "ok, I have the data, what interesting things can I do with it?"
Azeem: How did you hear about the class?
Bryan: I am pretty involved in the meet up scene. I go to many meet ups, and I saw one of Vivian's d3 classes. I happened to have been dabbling in d3 itself at the time, again, visualization. So I met her, and I learned about the school through her. I was finished building a data system at work and I wanted to start asking the serious questions: how can we make the advertising more effective, what can we learn from the data. And I thought I did not necessarily have all the tools for that. So that was the impetus.
Azeem: What did you think of the classes?
Bryan: I thought they were great. They are not everybody's cup of tea because you spend all day Sunday or Saturday, 8 hours, learning and coding and that can be a lot for some people. But that intense environment where you have to learn and think and code and put it all together and then go home and do homework, for a working professional, it is great. It really helped things click into place. It helped that I had these kinds of problems at work that I could apply to what I was learning.
Azeem: How was NYC Data Science Academy different from your academic training; because it seems like what we hear a lot is that academic training isn't enough for something like what Vivian is offering.
Bryan: Before I took the course I looked at a lot of the academic offerings and there were some very good schools offering masters degrees in quote unquote data science, and most of them were great, but there were two stumbling blocks. One was the price. Some of them were $60k a year for a two year program. That's out of reach for me. And I would have had to quit my job and do that full time. Also half the material these programs teach I already know. I just wanted the statistics, the math, the machine learning. With the master's programs you can't pick and choose what you need. Some of them won't even let you audit more than two courses before they require you to apply to the program.
Another stumbling block, was having to take GRE's and have good grades in college; the programs seemed to want me to prove that I could do this kind of high level work, even though I've already been doing a lot of it. I went back and forth because I would like the degree, I think that would look good on my resume.
But then I came back to NYC Data Science Academy. After going through the syllabus I knew that what Vivian is offering is just what I was looking for. Also, it offers flexibility in terms of what courses I take and in pricing. It felt like a smaller stepping stone, and one that fit better with my needs.
Azeem: Would you suggest to other people that they take a class; why and why not?
Bryan: Absolutely! There are many reasons. For one thing, back when I was learning computer science, which was as little as 10 years ago, machine learning was there but it wasn't as important as it is today. It's good to get your hands on the tools of the trade: R, python, visualization libraries, putting it all together, and on the web. Those are things Vivian teaches you and gives you the tools to use. I think that answers your question?
Azeem: How has taking the classes benefited you?
Bryan: Like I said, the class really scratched an itch. It allowed me to think at a higher level about the various data problems I have, and then concretely think about how to apply specific models. I was able to briefly use some of this knowledge at work. If I were still in my old position or trying to work for another start up in the area, they'd want to know what I did; what I did with neural networks, or regression, or classification problems. There are so many applications, and many companies that are definitely looking for that skill set. If you can explain really well what you learned or what you did in that class, that goes way further than saying what course you took in college.
Azeem: Tell me a little about what you are doing now in your start up.
Bryan: I have a start up based around sensor platforms. We built a small prototype that can easily be changed out from one problem type to another. Our current problem is location data; so a mobile location unit that is cheaper than a cell phone but also not likely to be taken by the driver. We are tracking fleets of vehicles, whether it be buses or taxis, for small to mid size businesses or cities. These small cities, with only a few buses, cannot afford the major packages from the major vendors. We are offering a more attractive solution in that space. We offer the analytics as well, for the sensor platform, at attractive prices.
Azeem: This last question is rather free form, just anything else you would like to add about your experience with the classes?
Bryan: I think I will add this. The thing I really enjoyed about Vivian's class is, she makes you go line by line to learn what you are doing, which is more than a lot of professors are willing to do or spend the time doing. It might seem like she is focused on R, but she is actually really focused on teaching the algorithms, and understanding why the algorithms work, and why certain choices are made by them. A lot of programmers don't necessarily get the luxury of playing around with algorithms in day to day work. So it is a nice break from the drudgery of every day work to wrap your mind around interesting concepts; fiddle with the data, visualize it, and really come to understand what you are working with. You are not just pressing a button and looking for the answer. And I think that really gave me confidence in what I am doing; and confidence to keep learning. That's the important part; you have to keep learning. You'll notice that Google just bought two deep learning companies; deep learning and neural networks are the big rage. These algorithms do everything from photo identification to learning about stocks. There are a lot of applications, and a lot of money in the start up space around these concepts. So this is all really relevant and a fascinating field.